DOCSLIB.ORG

POPL: G: Reasoning About Incentive Compatibility

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
POPL: G: Reasoning About Incentive Compatibility POPL: G: Reasoning about Incentive Compatibility Suguman Bansal Rice University [email protected] Abstract haviors that conform with agent objectives are called the solution We study equilibrium computation in regular repeated games, a concepts of the game. model of infinitely repeated, discounted, general-sum games where Algorithmic analysis of solution concepts in games has been agent strategies are given by a finite automaton. Our goal is to com- widely studied in algorithmic game theory [24]. In this field, com- pute, for subsequent querying and analysis, the set of all Nash equi- plex games are studied as repeated games [1, 14, 17]. An important libria in such a game. Our technical contribution is an algorithm solution concept is that of Nash equilibria [23]. Intuitively, Nash for computing all Nash equilibria in a regular repeated game. We equilibria is a state of a game in which no agent can increase its use our algorithm for automated analysis of properties such as in- reward by changing its own behavior only. centive compatibility on complex multi-agent systems with reward There is a wide literature on modeling repeated games as finite maximizing agents, such as the Bitcoin protocol. state machines [19, 27, 28]. The problem of computing Nash equi- libria in repeated games has received a lot of interest. The seminal work by Abreu, Pearce, and Stacchetti [2], and its followups [10], 1. Introduction considered the problem of computing all or some equilibrium re- Games [26], defined as multi-agent systems with reward maximiz- wards rather than equilibria. The problem of efficiently computing ing agents, find applications in modeling and understanding prop- a single equilibrium strategy in infinite games has been studied be- erties in a wide variety of fields, such as economics [11, 13, 20, fore [3, 15, 16]. There are also a few approaches to computing rep- 25, 32], social sciences [8, 21, 22], biology [4, 9, 18, 33], etc. resentations of approximations to the set of equilibria [5–7]. Except With the emergence of massive online protocols such as auctions, for work on finding one equilibrium, most other approaches do not adopted by cyber giants such as Google Auctions and eBay, games guarantee crisp complexity bounds. have generated much interest for utilitarian research from within For our purpose of analysis in repeated games, we are interested the computer science community. Presence of bugs in these proto- in the problem of computing all Nash equilibria in a game. In this cols have tremendous impact on both, the protocol providers and paper, we combine ideas from Game theory, and tools and tech- its large number of users. Hence, the analysis of properties on such niques developed in Programming languages and Formal methods systems is crucial. to realize this goal. First, we define a finite state machine based One concrete instance of the analysis of such protocols arises framework to model repeated games. We call our framework regu- in the context of the Bitcoin protocol. Bitcoins are one of the lar repeated games. Next, we develop an algorithm ComputeNash most widely used on-line currency in today’s time. The protocol to compute and enumerate all Nash equilibria in a regular repeated by which Bitcoins are mined (minted) is called the Bitcoin pro- game. Unlike previous approaches, our algorithm has a crisp com- tocol. Agents (miners) of the protocol receive rewards by mining plexity bound, and can determine all Nash equilibria in a regular more bitcoins. Earlier it was believed that miners receive the most repeated game. Lastly, as a proof of concept, we perform case stud- reward when they mine following the honest policies of the proto- ies on the Bitcoin protocol and repeated auctions, and corroborate col. Since the protocol rewarded miners proportional to the num- previously known results on them via analysis by ComputeNash. ber of bitcoins they mined, it was considered to be a fair proto- The merit in our approach over earlier approaches for the analysis col. However, a surprising result by Eyal and Sirer disproved this of games is that our analysis approach is automated, hence quicker misconception[12]. They proved via rigorous manual analysis that and not prone to errors. For the Bitcoin protocol case, we first ar- miners can receive greater rewards by following dishonest policies. gue that the Bitcoin protocol can be viewed as a repeated game, and A game is said to be incentive compatible if it ensures that reward hence we can apply ComputeNash for its analysis. maximizing agent will behave according to the honest policies of a This paper is organized as follows. Section 2 provides the nec- game. Hence, Eyal and Sirer proved that the Bitcoin protocol is not essary preliminary background from game theory and program- incentive compatible. ming languages. Section 3 introduces and defines regular repeated Incentive compatibility is a desirable property in games, since it games as a model for repeated games. Section 4 glosses over the ensures that all agents behave honestly in the game. However, erro- description of ComputeNash. Lastly, Section 5 demonstrates how neous system design is common. Therefore, analysis of properties ComputeNash can be used for automatic analysis of incentive such as incentive compatibility is critical for good game design. compatibly in the Bitcoin protocol and repeated auctions. Unfortunately, as demonstrated in the Bitcoin protocol case, even in today’s times most of this analysis is performed manually. One 2. Preliminaries cannot simply port results on verification of traditional systems in the context of games since their verification differs due to the pres- 2.1 Repeated Games ence of reward maximizing agents. More specifically, unlike in tra- Repeated games [17] are a model for multi-agent systems in which ditional systems, not all behaviors in a game need to be analyzed. agents interact with each other infinitely often. In each round of In the analysis of games, one can safely ignore behaviors that do interaction, each agent picks an action. These actions are picked not conform with the reward maximizing objectives of agents. Be- synchronously by all agents i.e. in each round of interaction, agents 1 2016/5/9 are not aware of what actions other agents will take in that round. 3. Regular repeated games (RRG) However, agents become aware of actions taken by other agents in We define Regular repeated games as a formal model for repeated all previous rounds. Agents receive a payoff (or reward) based on games. Regular repeated games builds on top of the well estab- the actions taken by each agent in an interaction. The payoff of an lished Rubinstein’s model for repeated games [27]. agent from a repeated game is computed by aggregating the rewards Formally, a regular repeated game G is given by a set of k received by the agent in each round using the discounted sum. The agents, where the i-th agent Pi consists of a finite set Action(i) discounted sum of a sequence of payoff A = {a0,a1,a2 . } of actions, and a finite set Strategy(i) of strategies. A strategy M with discount factor d > 1, denoted by DS(A,d), is given by ∞ i for Pi is B¨uchi automaton with weights. Figure 1 shows a strat- Σ ai/d . i=0 egy for the agent Pi. Each transition of the automaton represents Each agent in a repeated games has its own strategies. Intu- one round of the repeated game. Transitions occur over the alpha- itively, a strategy can be thought of as a guideline for the agent on bet of (a, (a)−i) where a ∈ Action(i) is the action of the agent how to pick an action in a round based on actions taken by all agents Pi, and (a)−i ∈ Action(j) is the environment action. Agent in previous rounds. For example, in an auction one strategy for an j=6 i Pi receives a reward along each transition. In Figure 1, transition agent is to bid truthfully, and another strategy is to never bid truth- (C,D),0 Q fully. All agents pick their strategy for the entire game only once in q0 −−−−→ q1 denotes that when the agent acts C, and the environ- the beginning. Agents pick their strategies synchronously. Strategy ment acts D, then the agent receives a reward of 0. profile, denoted by M, is a k-tuple of strategies (M1,..., Mk) Intuitively, accepting runs of a strategy M offer an agent-level i view of executions of the game. Specifically, consider a transition where M is the strategy of the i-th agent. ′ Various solution concepts have been defined as a means to char- (q, (a)−i,ai,p,q ) that appears at the j-th position of an accepting acterize which strategy an agent should pick in order to receive run of M. This means that there is a possible execution of the greater rewards from a game. This is important, since we assume game in which: (i) the i-th agent is at state q immediately before that agents in multi-agent systems are reward-maximizing by na- the j-th round of the game; (ii) in this round, the i-th agent and ture. Two such solution concepts are: Nash equilibria and best re- its environment synchronously perform the actions ai and (a)−i, respectively; (iii) the concurrent action leads agent Pi to transition sponse strategy. ′ Before defining these concepts, we introduce some necessary to state q at the end of this round, and receive payoff p. ′ notation. For each agent Pi and its strategy M , we define a strat- ′ Strategy profiles Now we define the semantics of interactions egy profile M[i := M ] obtained by starting with M, and then between agents P1,..., Pk that constitute a game G.
Load more

Recommended publications
  • Collusion Constrained Equilibrium
    Theoretical Economics 13 (2018), 307–340 1555-7561/20180307 Collusion constrained equilibrium Rohan Dutta Department of Economics, McGill University David K. Levine Department of Economics, European University Institute and Department of Economics, Washington University in Saint Louis Salvatore Modica Department of Economics, Università di Palermo We study collusion within groups in noncooperative games. The primitives are the preferences of the players, their assignment to nonoverlapping groups, and the goals of the groups. Our notion of collusion is that a group coordinates the play of its members among different incentive compatible plans to best achieve its goals. Unfortunately, equilibria that meet this requirement need not exist. We instead introduce the weaker notion of collusion constrained equilibrium. This al- lows groups to put positive probability on alternatives that are suboptimal for the group in certain razor’s edge cases where the set of incentive compatible plans changes discontinuously. These collusion constrained equilibria exist and are a subset of the correlated equilibria of the underlying game. We examine four per- turbations of the underlying game. In each case,we show that equilibria in which groups choose the best alternative exist and that limits of these equilibria lead to collusion constrained equilibria. We also show that for a sufficiently broad class of perturbations, every collusion constrained equilibrium arises as such a limit. We give an application to a voter participation game that shows how collusion constraints may be socially costly. Keywords. Collusion, organization, group. JEL classification. C72, D70. 1. Introduction As the literature on collective action (for example, Olson 1965) emphasizes, groups often behave collusively while the preferences of individual group members limit the possi- Rohan Dutta: [email protected] David K.
    [Show full text]
  • Informational Size and Incentive Compatibility
    Econometrica, Vol. 70, No. 6 (November, 2002), 2421–2453 INFORMATIONAL SIZE AND INCENTIVE COMPATIBILITY By Richard McLean and Andrew Postlewaite1 We examine a general equilibrium model with asymmetrically informed agents. The presence of asymmetric information generally presents a conflict between incentive com- patibility and Pareto efficiency. We present a notion of informational size and show that the conflict between incentive compatibility and efficiency can be made arbitrarily small if agents are of sufficiently small informational size. Keywords: Incentive compatibility, mechanism design, incomplete information. 1 introduction The incompatibility of Pareto efficiency and incentive compatibility is a central theme in economics and game theory. The issues associated with this incompatibility are particularly important in the design of resource allocation mechanisms in the presence of asymmetrically informed agents where the need to acquire information from agents in order to compute efficient outcomes and the incentives agents have to misrepresent that information for personal gain come into conflict. Despite a large literature that focuses on these issues, there has been little work aimed at understanding those situations in which informational asymmetries are quantitatively important. Virtually every transaction is characterized by some asymmetry of information: any investor who buys or sells a share of stock generally knows something rel- evant to the value of the share that is not known to the person on the other side of the transaction. In order to focus on more salient aspects of the problem, many models (rightly) ignore the incentive problems associated with informa- tional asymmetries in the belief that, for the problem at hand, agents are “infor- mationally small.” However, few researchers have investigated the circumstances under which an analysis that ignores these incentive problems will yield results similar to those obtained when these problems are fully accounted for.
    [Show full text]
  • Agency Business Cycles
    A Service of Leibniz-Informationszentrum econstor Wirtschaft Leibniz Information Centre Make Your Publications Visible. zbw for Economics Golosov, Michail Ju.; Menzio, Guido Article Agency business cycles Theoretical Economics Provided in Cooperation with: The Econometric Society Suggested Citation: Golosov, Michail Ju.; Menzio, Guido (2020) : Agency business cycles, Theoretical Economics, ISSN 1555-7561, Wiley, Hoboken, NJ, Vol. 15, Iss. 1, pp. 123-158, http://dx.doi.org/10.3982/TE3379 This Version is available at: http://hdl.handle.net/10419/217101 Standard-Nutzungsbedingungen: Terms of use: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Documents in EconStor may be saved and copied for your Zwecken und zum Privatgebrauch gespeichert und kopiert werden. personal and scholarly purposes. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle You are not to copy documents for public or commercial Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich purposes, to exhibit the documents publicly, to make them machen, vertreiben oder anderweitig nutzen. publicly available on the internet, or to distribute or otherwise use the documents in public. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, If the documents have been made available under an Open gelten abweichend von diesen Nutzungsbedingungen die in der dort Content Licence (especially Creative Commons Licences), you genannten Lizenz gewährten Nutzungsrechte. may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by-nc/4.0/ www.econstor.eu Theoretical Economics 15 (2020), 123–158 1555-7561/20200123 Agency business cycles Mikhail Golosov Department of Economics, University of Chicago and NBER Guido Menzio Department of Economics, New York University and NBER We develop a theory of endogenous and stochastic fluctuations in economic ac- tivity.
    [Show full text]
  • Political Game Theory Nolan Mccarty Adam Meirowitz
    Political Game Theory Nolan McCarty Adam Meirowitz To Liz, Janis, Lachlan, and Delaney. Contents Acknowledgements vii Chapter 1. Introduction 1 1. Organization of the Book 2 Chapter 2. The Theory of Choice 5 1. Finite Sets of Actions and Outcomes 6 2. Continuous Outcome Spaces* 10 3. Utility Theory 17 4. Utility representations on Continuous Outcome Spaces* 18 5. Spatial Preferences 19 6. Exercises 21 Chapter 3. Choice Under Uncertainty 23 1. TheFiniteCase 23 2. Risk Preferences 32 3. Learning 37 4. Critiques of Expected Utility Theory 41 5. Time Preferences 46 6. Exercises 50 Chapter 4. Social Choice Theory 53 1. The Open Search 53 2. Preference Aggregation Rules 55 3. Collective Choice 61 4. Manipulation of Choice Functions 66 5. Exercises 69 Chapter 5. Games in the Normal Form 71 1. The Normal Form 73 2. Solutions to Normal Form Games 76 3. Application: The Hotelling Model of Political Competition 83 4. Existence of Nash Equilibria 86 5. Pure Strategy Nash Equilibria in Non-Finite Games* 93 6. Application: Interest Group Contributions 95 7. Application: International Externalities 96 iii iv CONTENTS 8. Computing Equilibria with Constrained Optimization* 97 9. Proving the Existence of Nash Equilibria** 98 10. Strategic Complementarity 102 11. Supermodularity and Monotone Comparative Statics* 103 12. Refining Nash Equilibria 108 13. Application: Private Provision of Public Goods 109 14. Exercises 113 Chapter 6. Bayesian Games in the Normal Form 115 1. Formal Definitions 117 2. Application: Trade restrictions 119 3. Application: Jury Voting 121 4. Application: Jury Voting with a Continuum of Signals* 123 5. Application: Public Goods and Incomplete Information 126 6.
    [Show full text]
  • Lectures in Contract Theory1
    Lectures in Contract Theory1 Steve Tadelis and Ilya Segal2 UC Berkeley and Stanford University Preliminary and Incomplete December 2005 1These notes were prepared for a second year graduate course in Contract theory at Stanford University (Econ 282/291) and UC Berkeley (Econ 206). They are preliminary and incomplete. Thanks also to Debbie Johnston for typing the first version and to Scott Hemphill for excellent research assistance. 2Copyright c 1998-2005 by Steven Tadelis and Ilya Segal. All rights reserved. No part of these° notes may be reproduced in any form by any electronuic or mechanical means without writen permission from the author. Contents IIntroduction 5 0.1Tradewithsmallnumbersofagents............... 6 0.2 Incentives ............................. 7 0.3BoundedRationality....................... 8 0.4Contracts,Mechanisms,Institutions............... 8 II Hidden Information 10 1 The Principal-Agent Model 12 1.1Setup................................ 12 1.1.1 Preferences........................ 12 1.1.2 Contracting........................ 13 1.2Single-CrossingandtheMonotonicityofChoice........ 13 1.3TheFullInformationBenchmark................ 16 1.4TheRevelationPrinciple..................... 18 1.5SolutionwithTwoTypes..................... 20 1.5.1 ManyDiscreteTypes................... 24 1.6SolutionwithaContinuumofTypes.............. 25 1.6.1 TheRelaxedProblem................... 28 1.6.2 Whattodoifmonotonicitybinds(technical)...... 31 1.7ApplicationsoftheModel.................... 32 1.7.1 SecondDegreePriceDiscrimination..........
    [Show full text]
  • Mechanisms for Making Crowds Truthful
    Journal of Arti¯cial Intelligence Research 34 (2009) 209-253 Submitted 06/2008; published 03/2009 Mechanisms for Making Crowds Truthful Radu Jurca [email protected] Google Inc., Switzerland Boi Faltings [email protected] Ecole Polytechnique F¶ed¶erale de Lausanne (EPFL) Arti¯cial Intelligence Laboratory (LIA) CH-1015 Lausanne, Switzerland Abstract We consider schemes for obtaining truthful reports on a common but hidden signal from large groups of rational, self-interested agents. One example are online feedback mechanisms, where users provide observations about the quality of a product or service so that other users can have an accurate idea of what quality they can expect. However, (i) providing such feedback is costly, and (ii) there are many motivations for providing incorrect feedback. Both problems can be addressed by reward schemes which (i) cover the cost of obtaining and reporting feedback, and (ii) maximize the expected reward of a rational agent who reports truthfully. We address the design of such incentive-compatible rewards for feedback generated in environments with pure adverse selection. Here, the correlation between the true knowledge of an agent and her beliefs regarding the likelihoods of reports of other agents can be exploited to make honest reporting a Nash equilibrium. In this paper we extend existing methods for designing incentive-compatible rewards by also considering collusion. We analyze di®erent scenarios, where, for example, some or all of the agents collude. For each scenario we investigate whether a collusion-resistant, incentive-compatible reward scheme exists, and use automated mechanism design to specify an algorithm for deriving an e±cient reward mechanism.
    [Show full text]
  • A Strong Minimax Theorem for Informationally-Robust Auction Design∗
    A Strong Minimax Theorem for Informationally-Robust Auction Design∗ Benjamin Brooks Songzi Du October 9, 2020 Abstract We study the design of profit-maximizing mechanisms in environments with in- terdependent values. A single unit of a good is for sale. There is a known joint distribution of the bidders' ex post values for the good. Two programs are consid- ered: (i) Maximize over mechanisms the minimum over information structures and equi- libria of expected profit; (ii) Minimize over information structures the maximum over mechanisms and equi- libria of expected profit. These programs are shown to have the same optimal value, which we term the profit guarantee. In addition, we characterize a family of linear programs that relax (i) and produce, for any finite number of actions, a mechanism with a corresponding lower bound on equilibrium profit. An analogous family of linear programs relax (ii) and produce, for any finite number of signals, an information structure with a corresponding upper bound on equilibrium profit. These lower and upper bounds converge to the profit guarantee as the numbers of actions and signals grow large. Our model can be extended to allow for demand constraints, multiple goods, and ambiguity about the value distribution. We report numerical simulations of approximate solutions to (i) and (ii). Keywords: Mechanism design, information design, optimal auctions, profit maximiza- tion, interdependent values, max-min, Bayes correlated equilibrium, direct mechanism. JEL Classification: C72, D44, D82, D83. ∗Brooks: Department of Economics, University of Chicago, [email protected]; Du: Department of Economics, University of California San Diego, [email protected]. This research has been supported by NSF #1757222.
    [Show full text]
  • Mechanism Design and the Revelation Principle CSCI 1440/2440 2021-02-3
    Mechanism Design and the Revelation Principle CSCI 1440/2440 2021-02-3 First, we review the mechanism design formalism, in terms of games of incomplete information. Then, we introduce the notion of an indirect mechanism, presenting the English and Dutch auctions as examples. Finally, we cover the Revelation Principle,1 which can transform any 1 Roger B Myerson. Optimal auction mechanism into a direct one, ensuring incentive compatibility. design. Mathematics of operations research, 6(1):58–73, 1981; and Roger B Myerson. Incentive compatibility and 1 Mechanism Design Framework the bargaining problem. Econometrica: journal of the Econometric Society, pages 61–73, 1979 Mechanism design has been referred to as the engineering branch of game theory. It is concerned with designing mechanisms (i.e., games) such that the outcomes that arise when the games are played by rational agents (i.e., the equilibria) achieve some desiderata. The mechanism design framework thus consists of three parts: the mech- anism formalism, which is slightly generalizes games of incomplete information; solution concepts, or equilibria, which serve to predict the outcome of the mechanism/game; and desiderata, or objectives. Mechanisms The interaction between the designer of a mecha- nism and its participants can be modeled as a multi-stage game. The designer moves first by selecting a mechanism. The partici- pants observe the mechanism, and move thereafter. We restrict our present attention to a two-stage game, in which the participants play a simultaneous-move (i.e., one-shot) game in the second stage. After the designer announces their choice of mechanism, the agents face a game of incomplete information.
    [Show full text]
  • Dynamic Reserve Prices for Repeated Auctions: Learning from Bids?
    Dynamic Reserve Prices for Repeated Auctions: Learning from Bids? Yash Kanoria Hamid Nazerzadeh Columbia University University of Southern California July 2014 Abstract. A large fraction of online advertisement is sold via repeated second price auctions. In these auctions, the reserve price is the main tool for the auctioneer to boost revenues. In this work, we investigate the following question: Can changing the reserve prices based on the previous bids improve the revenue of the auction, taking into account the long-term incentives and strategic behavior of the bidders? We show that if the distribution of the valuations is known and satisfies the stan- dard regularity assumptions, then the optimal mechanism has a constant reserve. However, when there is uncertainty in the distribution of the valuations, previous bids can be used to learn the distribution of the valuations and to update the reserve price. We present a simple, ap- proximately incentive-compatible, and asymptotically optimal dynamic reserve mechanism that can significantly improve the revenue over the best static reserve. 1 Introduction Advertising is the main component of monetization strategies of most Internet companies. A large fraction of online advertisements are sold via advertisement exchanges platforms such as Google's Doubleclick (Adx) and Yahoo!'s Right Me- dia.1 Using these platforms, online publishers such as the New York Times and the Wall Street Journal sell the advertisement space on their webpages to adver- tisers. The advertisement space is allocated using auctions where advertisers bid in real time for a chance to show their ads to the users. Every day, tens of billions of online ads are sold via these exchanges [Muthukrishnan, 2009, McAfee, 2011, arXiv:2002.07331v1 [cs.GT] 18 Feb 2020 Balseiro et al., 2011, Celis et al., 2014].
    [Show full text]
  • The Implementation of Social Choice Rules: Some General Results on Incentive Compatibility
    The Implementation of Social Choice Rules: Some General Results on Incentive Compatibility Partha Dasgupta; Peter Hammond; Eric Maskin The Review of Economic Studies, Vol. 46, No. 2. (Apr., 1979), pp. 185-216. Stable URL: http://links.jstor.org/sici?sici=0034-6527%28197904%2946%3A2%3C185%3ATIOSCR%3E2.0.CO%3B2-E The Review of Economic Studies is currently published by The Review of Economic Studies Ltd.. Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/about/terms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/journals/resl.html. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. The JSTOR Archive is a trusted digital repository providing for long-term preservation and access to leading academic journals and scholarly literature from around the world. The Archive is supported by libraries, scholarly societies, publishers, and foundations. It is an initiative of JSTOR, a not-for-profit organization with a mission to help the scholarly community take advantage of advances in technology. For more information regarding JSTOR, please contact [email protected].
    [Show full text]
  • Ambiguity Aversion Solves the Conflict Between Efficiency and Incentive
    Ambiguity aversion solves the conflict between efficiency and incentive compatibility∗ Luciano De Castroy Nicholas C. Yannelisz First Version: December 2008 This Version: November 2010 Abstract The conflict between Pareto optimality and incentive compatibility, that is, the fact that some Pareto optimal (efficient) allocations are not incen- tive compatible is a fundamental fact in information economics, mechanism design and general equilibrium with asymmetric information. This impor- tant result was obtained assuming that the individuals are expected utility maximizers. Although this assumption is central to Harsanyi’s approach to games with incomplete information, it is not the only one reasonable. In fact, a huge literature criticizes EU’s shortcomings and propose alternative preferences. Thus, a natural question arises: does the mentioned conflict extend to other preferences? We show that when individuals have (a special form of) maximin expected utility (MEU) preferences, then any efficient al- location is incentive compatible. Conversely, only MEU preferences have this property. We also provide applications of our results to mechanism de- sign and show that Myerson-Satterthwaite’s negative result ceases to hold in our MEU framework. Keywords: Asymmetric information, ambiguity aversion, Incentive com- patibility, mechanism design, first-best, second-best. JEL Codes: D50, D81, D82. ∗We are grateful to Marialaura Pesce, Marciano Siniscalchi and Costis Skiadas for helpful conversations. y Department of Managerial Economics and Decision Sciences, Kellogg School of Manage- ment, Northwestern University, Evanston IL 60208. zDepartment of Economics, University of Illinois at Urbana-Champaign and Economics- School of Social Sciences, The University of Manchester. E-mail: [email protected]. 1 Contents 1 Introduction3 2 Preliminaries6 2.1 Goods and allocations........................6 2.2 Allocations and endowments....................9 2.3 Incentive compatibility......................
    [Show full text]
  • Information Revelation in Competing Mechanism Games
    Information Revelation in Competing Mechanism Games Andrea Attar¤ Eloisa Campioni† Gwenaël Piaser‡ November 22, 2011 Abstract We consider multiple-principal multiple-agent games of incomplete information. In this context, we identify a class of direct and incentive compatible mechanisms: each principal privately recommends to each agent to reveal her private information to the other principals, and each agent behaves truthfully. We show that there is a rationale in restricting attention to this class of mechanisms: If all principals use direct incentive compatible mechanisms that induce agents to behave truthfully and obediently in every continuation equilibrium, there are no incentives to unilaterally deviate towards more sophisticated mechanisms. We develop examples to show that private recommendations are a key element of our construc- tion, and that the restriction to direct incentive compatible mechanisms is not sufficient to provide a complete characterization of all pure strategy equilibria. Key words: Incomplete information, competing mechanisms, information revelation. JEL Classification: D82. ¤University of Rome II, Tor Vergata and Toulouse School of Economics (IDEI-PWRI); [email protected]. †University of Rome II, Tor Vergata; [email protected]. ‡IPAG Business School, Paris; [email protected]. 1 Introduction This paper contributes to the analysis of competing mechanism games with incomplete infor- mation. Specifically, we consider a scenario in which several principals simultaneously design incentive contracts in
    [Show full text]